1. Quality of Consumer Services: US-Japan Comparison Based on a Survey of the Willingness to Pay
The 3rd Hitotsubashi Summer Institute
The Fourth Asia KLEMS Conference
July 31, Shoin Hall, Hitotsubashi University
Kyoji Fukao (Hitotsubashi University)
Kenta Ikeuchi (RIETI)
Miho Takizawa (Toyo University)

2. Motivation
Preceding studies have shown that labor productivity in Japan’s service sector is only about one half of the US level. These studies are based on PPPs of output and intermediate input.
Jorgenson, Nomura, and Samuels 2016, METI White Paper (2013, based on EU KLEMS PPP)
As data for the US-Japan PPP estimation for service output and input, many studies use the results of the ICP and government statistics such as METI’s Survey on Foreign and Domestic Price Differentials for Industrial Intermediate Input.
However, these data do not fully take service quality into account
For instance, in the case of rail transportation in cities, the ICP does not take account of the frequency of delays, frequency of trains, cleanness, number of crimes in trains and stations, etc.
The Japan Productivity Center and Hitotsubashi University have jointly conducted an internet survey on service quality.
First, we explain the current status of the productivity in Japan by comparing the level and the dynamics of industry-level labor productivity, which is defined as the ratio of real value-added to total hours worked, between Japan and the U.S.

3. 1. Labor Productivity 101 Productivity
　　　　＝ Output (e.g., value added) ÷ Input (e.g., total hours worked)
Aging & declining population
⇒ Need to improve productivity!!!

4. 2. Japan & US Labor Productivity?
Typical illustration:
Japan/US(=100) in period
Δof Japan/US(=100) from to
Step back:
Decomposition of Δ of Japan from 1997 to 2010
Decomposition of Δ of Japan/US(=100) from 1997 to 2010
Further decomposition Δ of Japan in subperiods
Data: (Japan) JIP2015, (US) World KLEMS, BEA & BLS,
(PPP) EU KLEMS.
Not quality-adjusted
Unfortunately, we only have the data up to 2012…

5. ※Takizawa and Japan Productivity Center(2016).
　 Source: RIETI, JIP Database; WORLD-KLEMS Database; BEA, Gross Domestic Product (GDP) by Industry Data; BLS, Labor Productivity and Costs; PPP: EU－KLEMS Database.　　　　

6. Note: Timmer and Inklaar (2008)
Only market economy (average) in 2005
Source: Author’s calculations on GGDC Productivity Level Database.

7. Δ of Japan/US(=100)
Japan/US Gap↓
Japan/US Gap↑

8. Productivity improvement pattern ① to ⑤
＝ Output (e.g., value added) ÷ Input (e.g., total hours worked)
5 major patterns:
“Aggressive”
“Passive”
Productivity improvement pattern ① to ⑤
Output
Input １ ２ ３ ４ ５
“Efficient”

10. Efficient
Aggressive
Passive
Shrinking

11. Efficient

12. Aggressive
Passive

13. Shrinking

14. Relatively aggressive
Relatively shrinking

17. 3. Summary of the introduction
Not quality-adjusted labor productivity of Japan & US
Japan/US<1
Gap↑ in service sector
In many Japanese industries LP↑ if we don’t benchmark it relative to US.
If we benchmark, only in a few Japanese industries LP↑…
Industries have experienced heterogeneous paths in terms of VA/L
How does quality-adjustment affect these findings???

18. Outline of the survey for comparison of US-Japan differences in quality of services
Main questions
Differences in perceived quality of services between the US and Japan.
The US-Japan relative willingness to pay (WTP) is evaluated for 28 types of consumer services.
Survey participants
Japan: 480 Japanese who have stayed in the US for 3 or more months in the last 5 years.
US : 412 Americans who have stayed in Japan for 1 month or more in the last 5 years.
Military personnel were excluded.
Extracted from the monitors of Macromill, Inc.
Survey method
Internet survey.
Survey period
28 February to 21 March 2017 for the survey on Japanese.
14 March to 11 April 2017 for the survey on Americans.

19. Outline of the survey for comparison of US-Japan differences in quality of services (contd.)
Main Question: Suppose services of average Japanese quality were offered in the US in English. If the Japanese service was better in quality than the corresponding US service, how much more would you be willing to pay for the Japanese service? Conversely, if the Japanese service was worse in quality, how much cheaper would it have to be for you to choose it over the corresponding US service?

20. Respondents by sex and age group
The observations are stratified by sex and age.
The weights for aggregation are calculated using the official demographics of each country (Ministry of Internal Affairs and Communications and US Census Bureau).
Japan survey
US survey
Sex
Age group
Population
(10 thousand)
# of valid
respondents
Male
20-29
640 36
2,340 67
30-39
773 49
2,180 61
40-49
960 57
2,020 62
50-59
775
2,129 21
60 or older
887
1,760 6
Female
609 37
2,231 72
750 59
2,148 66
938
2,047 40 58
2,224 11
932 25
1,942

21. Main Results

22. Main Results (contd.)

23. Next steps on service quality issues
Using Heckman's two-step estimation, we will make corrections for selection bias. For some industries, such as railroad and retail services, we will assess the economic value of each service characteristic using objective measures of service quality and price data and applying the hedonic approach. Based on these results, we will then estimate service quality differences.